Currently, when a follower's log is rolled back, the leader will crash again and again, which leads to no stable leader.
We plan to fix the issue. The expected behavior is:
- The cluster runs as normal. There is a stable leader.
- Leader will send log entries to the follower once its data is recovered.
Currently, when a follower's log is rolled back, the leader will crash again and again, which leads to no stable leader.
We plan to fix the issue. The expected behavior is: